KTHNOLOGT OF THE niDO-FACIFIC ISLANDS.. 



143 



Further lif?ht will be thrown on the hiitory of the Chmo-Hiraaliiic 

 nuiuerah vvht-u we comtuiro thtui aysteuiatically with th« tkytiiic, 

 C{iucju-*iiin md Africim. Out our esaini nation of the Chino-IIiraalaic. 

 Sonaitic and Afrk-nn. must be followed up by thut of the Scythic; and 

 Caucii:»ian, betbiv we can enter on Huch li com|mi'iijoiii. All the nu- 

 meral aystems of the Old World are more «r less connected, and pyiJit 

 to the diffdision of it^ earUt^r nrtsi and civilisation, or of the race« 

 theTn-Sflvea, from one family. For example both the prinimy and 

 secondary forma of rlip Chino-Himaliuc unit are found in the other Ea^t 

 AaimUc and the conneut-ed Westerik nystemii. The most impurtant of the 

 lecoitdary fornix ^Lini, Hum, 8ap is !to widely diffused that^ on &n 

 examination ot" a few numemlH:, it raig'ht be t!iouf:ht probable that it was* 

 CT-oIved from kam, kai>, at a very remote period utid before any extensive 

 dispersion of the ai^cient Asiatic j^ystem. Both the priiniiry and second- 

 ary formfi certininly api>p^ir to hiive been circulated to-if ttiur over other 

 provinc«i!i be-iides the Himaluio. But a compai-itton of all the Old World 

 system^a lead-( to the inference that the conneution of the Cbino-Himakic: 

 with thp Siiythic and more dintant numemlB is throujrh the primary 

 guttural f(jrui.s, aiid that the recurrence of identical secondary formi 

 and types in difft^rent provinces in ^feneral, attriUaUble to the «arae 

 cycle of t»honetie chaiig-e havinjf been independently repeated in each family 

 of iangua;fe. There are doubtlesii examples of a traii.-^ler of later forma 

 from one fiimily to another. But it retjiiirea strong evidence to e.stabli?ih 



in a single dialect. If we were to aftsunie that, at one time and while the 

 name-* atill remaine<l rompound, a diniect eiiHted in which the primary 

 form of the unit had been lost and *'am, sap was ita oaly currant for the 

 triual :4ystem might he thus presented. 



1st sirks. 

 1 mm 

 II nga 

 III 2| I nga-aarn 



2d srricfi. _ 

 jy ,2, 1 nufa.sam-sara ~ 



y fiij 2 ngfl.sam-njra 

 Vi , JJ, a ( or 5, 1 ) nga.aam-ngii*^'^!^ 



Vll 0, 1 ng-a.Aftm^npi.Jsrtni—sam 



In all the compoundg ef the 2d and M verier*, the hwt word would be 

 the distinctive and nermaaont one. The ofhce of the liri^t common rerm 

 of each series would lie merely to mark the seiios, and when a aliiht 

 phonetic chuc]j4;e wax induced in the la«t, tlii^ would ittM^tf ^tndice t^> murk 

 the serie:^, and tin* tirsi term would b»>conie a needle^» incninbrance. 

 Thus if, fixtm the action of the preceding- eifnienift of sound, o>r bv atrqiii* 

 aition tVuin anorlicr diiilecr, the nainf^ of 1 bKatmo saritj or .^ak, ihe^am uf 

 3 woald lie eniilileil to *h^il♦'nri^^ with iho n;^.i- not oidy m -i but throu^ih- 

 out thy liii^h'T seriti,H uf irtfhich it. f ti in^ fbe ratlix. Su, if the distinctive 

 or tiiiul .-sjimof^ iiocame sum, su or .*i, the initial n^fi-.-^aui, tir ita remnant 

 »atn, inl;rhF b" rej-'et«d. I'l I ke ni iiiinir ti t li.uMri} of nja 2 lo mjiji *vould 

 eJUil/lti the di.stuicUve nj^ of 5 to rejiJt;t the prepostid ttriu or Us remiiaut. 



